EcoSimRPresentation_Fin.Rmd

---
title: "EcoSimR Presentation"
author: "Morgan Southgate"
date: "April 18, 2017"
output: 
  html_document: 
    theme: cerulean
---
# EcoSimR - Null Model Analysis of Ecological Data

### The following code demonstrates how to use the EcoSimR package to conduct a species co-occurrence analysis using null models. The results of this analysis reveal if there are statistically significant patterns of species aggregation or segregation in the dataset. 

### The following code uses associated species data collected from 19 survey patches across five serpentine outcrops in northern VT, which was collected as part of a maidenhair fern habitat study in the summer of 2016. The following code transforms the associated species data into a presence-absence matrix that is then used for the null model analysis.

[Associated Species Data](AssociatedSppData_Serp.csv)


## 1. Create presence-absence matrix from species list data
```{r, echo=TRUE, message=FALSE, warning=FALSE}
# read in associated species data 
sppDat <- read.table("AssociatedSppData_Serp.csv",header=TRUE,sep=",",stringsAsFactors = FALSE)
head(sppDat)
```

```{r, echo=TRUE, message=FALSE, warning=FALSE}
# reshape data using dcast function in reshape2 package
library(reshape2)
PA <- dcast(sppDat,formula=SpeciesName~SitePatch)
head(PA)
dim(PA)
```


## 2. Conduct Species Co-occurrence Analysis using EcoSimR
### SIM9 null model algorithm- row and column sums fixed
#### CHECKER index
```{r, echo=TRUE, message=FALSE}
library(EcoSimR)
library(MASS)
# Run null model with SIM9 algorithm & CHECKER index
adMod1 <- cooc_null_model(PA,
                    algo="sim9",
                    metric="checker",
                    nReps=1000,
                    suppressProg=T)

# Summary and plots
summary(adMod1)
mean(adMod1$Sim)
plot(adMod1,type="hist")
plot(adMod1,type="cooc")
plot(adMod1,type="burn_in")
```

#### C score index
```{r, echo=TRUE, message=FALSE, warning=FALSE}
## Run null model with SIM9 algorithm and C score index
adMod2 <- cooc_null_model(PA,
                          algo="sim9",
                          metric="c_score",
                          nReps=1000,
                          suppressProg = T)

# Summary and plots
summary(adMod2)
plot(adMod2,type="hist")
plot(adMod2,type="cooc")
plot(adMod2,type="burn_in")
```

#### COMBO index
```{r, echo=TRUE,eval=F}
# Run null model with SIM9 algorithm and COMBO index
adMod3 <- cooc_null_model(PA,
                          algo="sim9",
                          metric="species_combo",
                          nReps=1000,
                          suppressProg = T)

# Summary and plots
summary(adMod3)
plot(adMod3,type="hist")
plot(adMod3,type="cooc")
```

### SIM2 null model algorithm- row sums fixed, columns equiprobable

#### CHECKER Index
```{r, echo=TRUE}
# Run null model with SIM2 algorithm and C score index
adMod4 <- cooc_null_model(PA,
                          algo= "sim2",
                          metric="c_score",
                          nReps=1000,
                          suppressProg=T)

# Summary and plots
summary(adMod4)
plot(adMod4,type="hist")
plot(adMod4,type="cooc")

```

#### C score Index
```{r, echo=TRUE}
# Run null model with SIM2 algorithm and CHECKER index
adMod5 <- cooc_null_model(PA,
                          algo= "sim2",
                          metric="checker",
                          nReps=1000,
                          suppressProg=T)

# Summary and plots
summary(adMod5)
plot(adMod5,type="hist")
plot(adMod5,type="cooc")

```


#### V ratio Index
```{r, echo=TRUE}
# Run null model with SIM2 algorithm and V ratio index
adMod6 <- cooc_null_model(PA,
                          algo= "sim2",
                          metric="v_ratio",
                          nReps=1000,
                          suppressProg=T)

# Summary and plots
summary(adMod6)
plot(adMod6,type="hist")
plot(adMod6,type="cooc")
```

#### COMBO Index
```{r, echo=TRUE}
# Run null model with SIM2 algorithm and COMBO index
adMod7 <- cooc_null_model(PA,
                          algo="sim2",
                          metric="species_combo",
                          nReps=1000,
                          suppressProg=T)

# Summary and plots
summary(adMod7)
plot(adMod7,type="hist")
plot(adMod7,type="cooc")
```